Where are Claude Code subagents stored?

Project-specific agents are stored in .claude/agents/, global user agents in ~/.claude/agents/, and plugin agents within the plugin's agents/ directory.

Can I restrict which tools a subagent can use?

Yes, you can use the 'tools' field in the YAML frontmatter to provide a comma-separated list of specific tools like Read, Grep, or Bash, following the principle of least privilege.

What is the difference between a project agent and a user agent?

Project agents are scoped only to the current directory and have the highest priority, while user agents are available across all projects on your system but have lower priority if names conflict.

Can I define subagents without creating a file?

Yes, you can define subagents dynamically using the --agents CLI flag with a JSON configuration for quick testing or session-specific tasks.

What models can I assign to a subagent?

Subagents can be configured to use specific model aliases like 'sonnet', 'opus', or 'haiku', or they can 'inherit' the model used in the main conversation.

Claude Code Agent Development

Name: Claude Code Agent Development
Author: astrosteveo

byastrosteveo

0•

开发者工具

Streamlines the creation and configuration of specialized AI subagents to automate complex developer workflows within Claude Code.

The Agent Development skill provides a comprehensive framework for building, managing, and optimizing subagents—specialized AI assistants that Claude Code can delegate tasks to. It standardizes the subagent creation process by providing clear guidance on YAML frontmatter configuration, file location priority, and system prompt engineering. Whether you are building a custom code reviewer, a dedicated debugger, or a test automation specialist, this skill ensures your subagents are configured with the correct tools, models, and permissions to operate effectively within their own context windows.

主要功能

01Support for project-level, user-level, and plugin-based agent scopes

02Best practices for designing effective subagent system prompts

030 GitHub stars

04Comprehensive guidance on tool access and permission mode configuration

05Standardized Markdown and YAML frontmatter templates for subagents

06Documentation for dynamic CLI-based agent definitions using JSON

使用场景

01Building a dedicated debugging subagent with access to specific shell and grep tools

02Creating a specialized code review agent that proactively checks for security vulnerabilities

03Developing automated test runners that trigger immediately after code modifications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add astrosteveo/claude-code-plugins agent-development

For use in Claude.ai and ChatGPT

主要功能

01Support for project-level, user-level, and plugin-based agent scopes

02Best practices for designing effective subagent system prompts

030 GitHub stars

04Comprehensive guidance on tool access and permission mode configuration

05Standardized Markdown and YAML frontmatter templates for subagents

06Documentation for dynamic CLI-based agent definitions using JSON

使用场景

01Building a dedicated debugging subagent with access to specific shell and grep tools

02Creating a specialized code review agent that proactively checks for security vulnerabilities

03Developing automated test runners that trigger immediately after code modifications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add astrosteveo/claude-code-plugins agent-development

For use in Claude.ai and ChatGPT